Ai powered service design thinking

After a few discussions with the Co-op design team around how many people are already exploring using Ai as part of their creative process at work, and with an hour to play with whilst on my day off, I finally joined the hype train.

I’ve been meaning to write something about the relationship between Service-dominant logic and service design for a while. It’s not a totally unexplored topic, but primarily within academic and die-hard service nerd circles. It’s always good to find your people!

The other nudge was that Notion have now included Ai as part of their note taking tool, so it was immediately accessible to me to try. So I did.

I thought I’d share the blog posts it generated, as well as a few reflections on the process and its use.

My first attempt at a prompt (although heavily informed by a lot of other prompts I’ve seen shared around recently) was:

Write a blog post about the relevance of Service-dominant logic to service design practice, targeted at designers working on complex services that involve software, physical products and processes

I’ve published the blog post it generated for you to read. Worth saying this was the first attempt, I didn’t ask it to make it longer or shorter, and I didn’t edit it at all.

My general reflections are:

On the positive side:

• Unsurprisingly it seems to have a decent grasp on both concepts and I’d suggest it highlights the most relevant parts of SDL to relate to service design
• It does appear to have differentiated between co-creation of value and co-creation as a design research method, which I had wondered if it would trip it up
• I don’t think there’s anything I would disagree with, it’s not dropped any clangers like I’ve seen other Ai do…like citing non-existent writers or articles etc.

On the less positive side:

• It’s lacking any practical application, it connects the theories of service design and service dominant logic, but little more, it’s just quite generic really
• Related, it reads like a post written to get clicks not provide meaningful value to the reader. Which I guess is mostly a reflection on the blog posts the Ai has been trained on
• Rather than try and make the content relevant to the target audience in subtle/nuanced ways it just refers to the importance to the group of people I’ve described
• Its quite cold and lacks any personality or humour

Based on that first attempt, I expanded on the original prompt:

Write a blog post about the relevance of service-dominant logic to service design practice, targeted at designers working on designing complex services made up of software, physical products and processes. The post should go beyond basic definitions of the different practice and draw on practical examples to help designers use the frameworks, whilst including elements of humour.

You can read the unedited blog post generated from that prompt too.

Some positive reflections on that post:

• Again, unsurprisingly the Ai responded to my updated prompt, weaving in aspects of things I asked for
• Overall it feels a much more meaningful blog post, aiming to actually help the reader rather than just get the clicks from SEO keywords
• It remains accurate and relevant, I’d feel comfortable sharing it with someone who asked for a summary of the links between the two concepts/frameworks
• Whilst not exactly a hilarious read, the humour added isn’t totally out of place, linking the end of the post to an earlier example (coffee machine)

On the less positive side:

• It still lacks any external references, which for me are an essential part of a credible blog post, it would feel strange to try and write anything meaningful about a topic without linking out or even referencing other authors/thinkers in related fields. I know plagiarism is one of the big concerns with Ai and the fact that the developers of SDL Vargo and Lusch are not even mentioned is odd here
• There’s something about it that still feels missing around practical application, harder to put my finger on though. But e.g. ‘how might I actually incorporate SDL into a service design process’ still isn’t really answered.

Of course, these are probably all explained by the prompt i.e. if I asked it to include references and make it even more practical it probably would…let’s try.

Third prompt:

Write a blog post about the relevance of service-dominant logic to service design practice, targeted at designers working on designing complex services made up of software, physical products and processes (without using that exact wording). The post should go beyond basic definitions of the different practice and draw on practical examples to help designers use the frameworks in their day to day work, whilst including elements of humour. The blog post should reference other relevant authors or thinker’s work to support it’s suggestions

Again I’ve published the full post on it’s unedited form for you to read.

Positive reflections on this third and (for now) final try:

• The step by step breakdown of application is more what I was hoping for, it responded well to that request and made it much more actionable I’d say
• It’s picked up on the request not to be so explicit with the target audience which in my eyes is better stylistically

The not so positive:

• It’s outright ignored my request for relevant names or references to other authors or thinker’s. Maybe this is an intentional decision to limit its use for writing academic work, but I’d not asked for actual references (footnotes etc) but just name drops of the most relevant people. That is still missing and feels intended.
• This post gets tripped up on the co-creation confusion. I’d suggest that in point 2 of it’s practical application steps it conflates co-creation as a design research approach with co-creation of value (core tenet of SDL) without realising.
• The attempt at ‘humour’ is weak, and feels pretty formulaic having seen it use two very similar approaches twice now.